Cohesyve · Practice for candidates
Preparing for a Financial Analyst role? Start by finding out where you stand.
Import the actual job you're going for, run a simulation built from it, and get a scored report showing which of these areas you're weakest in — before you spend weeks studying the wrong thing.
3
free assessments
$0
no card required
Full
scored report every run
Every
question type included
3 free assessments · No card required · Pro from $16/mo
Overview
Financial analyst hiring is unusually practical. Somewhere in the process you will be handed a spreadsheet and asked to build, fix or explain something under a clock, and the rest of the process tests whether you can turn that output into a decision a non-financial reader will act on. Most candidates prepare for the technical stage and neglect the commentary, which is where the field usually separates. This page covers the stages, the topics worth revising and the questions that come up.
The hiring process, stage by stage
Application and CV screen
Screened on sector relevance, the systems you have used and whether your CV shows ownership of a recurring cycle — budgeting, forecasting or month-end reporting — rather than ad hoc analysis.
How to prepare: Rewrite each role bullet around the cycle you owned and the decision your work fed. Name the systems explicitly — the ERP, the planning tool, the BI layer — because screeners search for them.
Recruiter or hiring manager call
A 20-30 minute conversation covering your reporting line, the size and nature of the numbers you handled, notice period and motivation for moving.
How to prepare: Prepare a two-minute description of the business you support, not just your tasks: what it sells, what drives its margin, and which line you were closest to. Have a plain answer for why you are leaving.
Modelling or case exercise
A timed spreadsheet task — build a short forecast from stated drivers, extend an existing model, or audit one seeded with errors. Sometimes taken home, increasingly done live with a screen shared.
How to prepare: Practise timed builds with inputs, calculations and outputs on separate sheets. Rehearse the audit habits too: trace precedents, check totals against components, test sign conventions and period alignment.
Technical and commercial interview
Usually with the finance manager or FP&A lead. Covers how the three statements interact, how you set and defend assumptions, and how you explained an uncomfortable number to a budget holder.
How to prepare: Be able to walk a change through the P&L, balance sheet and cash flow out loud. Prepare two variance stories decomposed into volume, price and mix rather than described in aggregate.
Stakeholder or panel interview
Meetings with the budget holders you would support — a commercial director, an operations lead — testing whether they would trust and understand you.
How to prepare: Practise explaining a variance with no numbers on screen, in under ninety seconds, leading with the conclusion. Prepare questions about how forecasts are actually used in their decisions.
What to study
Three-statement mechanics
Interviewers use it as a fast integrity check. If you cannot trace a depreciation change or a working capital movement through all three statements, they assume the rest of your modelling is memorised rather than understood.
Forecasting and driver-based modelling
Nearly every exercise asks you to build something that can be flexed. Models with assumptions buried inside formulas score badly regardless of whether the arithmetic is right.
Variance analysis
The most common interview question in the role, and the one most candidates answer by restating the difference. Decomposition is the whole test.
Spreadsheet mechanics and data handling
Timed exercises punish slow, mouse-driven work. Accuracy that only holds at a leisurely pace is the most common reason strong candidates fail the exercise stage.
Written commentary for non-financial readers
The output of the role is not the model; it is the paragraph a budget holder reads. Panels increasingly mark a short written piece alongside the exercise.
Cohesyve for candidates
Practise a Financial Analyst assessment before the real one
Run the same AI job simulations companies use to evaluate applicants. You get a scored report showing where you're strong and where you're not, plus what to work on.
Your preparation plan
Four weeks out
Rebuild the technical base and understand the business you are applying to.
- •Work through statement linkages until you can narrate any change out loud without notes
- •Read the employer's latest published results or annual report and identify what drives their margin
- •Build one forecast from scratch, untimed, with a clean assumptions sheet
- •List the systems on the job advert you have not used and read their basic reporting structure
Two weeks out
Move from knowledge to timed execution.
- •Do two timed modelling exercises of 60-90 minutes and note where the clock beat you
- •Practise auditing a model you deliberately broke, listing errors systematically
- •Draft three variance explanations decomposed into volume, price and mix
- •Write a one-paragraph board commentary for each and cut it by a third
The final week
Rehearse the spoken layer and tighten your examples.
- •Rehearse your two-minute summary of the business you currently support
- •Practise explaining a variance verbally in ninety seconds, with no visual aid
- •Prepare six questions about their planning cycle, reporting line and systems
- •Check your exercise environment — spreadsheet version, screen share, second monitor
The day before and the day of
Reduce friction rather than add material.
- •Reread the job advert and map each requirement to one example you can tell
- •Do one short warm-up build the morning of, purely to get your hands moving
- •Keep a blank sheet ready for scratch calculations during the live exercise
- •Plan to spend the first five minutes of any exercise reading and structuring, not typing
Questions to expect
“Walk me through what happens to the three statements if depreciation increases by 10.”
What they're really asking: Do you understand the accounting mechanically, or have you memorised model templates without the underlying logic?
How to answer: Take them in order — P&L first including the tax effect, then cash flow with depreciation added back, then the balance sheet on both sides. Finish by confirming it balances. Say the numbers out loud rather than gesturing at the concept.
“Revenue grew but margin fell. How would you work out why?”
What they're really asking: Can you decompose a variance, or will you restate the observation and stop?
How to answer: Name the decomposition first — volume, price, mix, and cost per unit — then say which data you would pull to isolate each. Give a worked example from your own experience and state which effect dominated.
“How do you set an assumption when you have no data?”
What they're really asking: Do you invent a number and defend it, or do you build something transparent enough to be corrected?
How to answer: Describe triangulating from an analogue — a comparable product, a prior launch, an external benchmark — putting the assumption in a single visible cell, flagging it as low confidence, and running a sensitivity around it.
“Tell me about a time your analysis changed a decision.”
What they're really asking: Does your work reach anyone, or does it stop at a file no one opens?
How to answer: Use a specific decision, name the audience, describe what the analysis showed that was not obvious, and state what they did differently. Keep the technical detail brief and the consequence concrete.
“A budget holder disagrees with your forecast. What do you do?”
What they're really asking: Can you hold a position with a commercial stakeholder without either caving or becoming obstructive?
How to answer: Separate the two possibilities: they have information you lack, or they dislike the answer. Describe asking which assumption they disagree with, testing it in the model, and showing the impact rather than arguing about the total.
“How do you check your own work?”
What they're really asking: Modelling errors are quiet. Do you have a habit, or do you rely on being careful?
How to answer: Name concrete checks — totals reconciling to components, a sense-check line built into the model, comparing output to a rough independent estimate, and having a second pair of eyes on anything going to a board.
What to bring as evidence
A sanitised model you built, with assumptions clearly separated, that you can talk through screen by screen
One variance analysis where your decomposition changed what the business believed was happening
A short written commentary you produced for a non-financial audience
A description of a reporting or forecasting process you took over and improved, with the before and after
A list of systems you have worked in and what you actually did in each, not just that you touched them
A worked example of an error you caught before it reached a decision-maker
Common mistakes
Hardcoding assumptions inside formulas, which makes a technically correct model unusable in the interview follow-up.
Describing a variance instead of decomposing it into volume, price and mix.
Spending the whole exercise on the build and leaving no time for the written commentary that carries the marks.
Presenting a single-point forecast with no sensitivity around its weakest assumption.
Talking about the finance function in isolation, with no view of what the business sells or where it makes money.
Answering stakeholder questions in accounting vocabulary that the stakeholder does not use.
Questions to ask them
- ?
How long does the monthly close take, and how much of the reporting pack is still manual?
- ?
Who owns the forecast — finance, or the budget holders with finance consolidating?
- ?
How often does the forecast get reforecast, and what triggers a reforecast outside the cycle?
- ?
Which system is the source of truth for revenue, and how much reconciliation sits between it and the pack?
- ?
Who reads my output, and what decision do they make with it?
- ?
What has this team been asked to improve in the next twelve months?
Frequently Asked Questions
How advanced does my Excel need to be?
Clean rather than exotic. Correct referencing, consistent formulas across a row, lookups and pivots handle almost every exercise. Elaborate techniques rarely earn marks, and an inflexible model always loses them.
Will the modelling exercise be timed or taken home?
Both are common, and live exercises have become more frequent because take-homes are hard to attribute. Ask which format you will face when you confirm the interview; it changes how you prepare.
Do I need a qualification to be considered?
For most analyst roles a part-qualification or a numerate degree is enough, and studying towards a qualification is often welcomed. Fully qualified is more commonly required as you move towards controller and reporting roles.
How much do they care about industry experience?
More than candidates expect for revenue-facing roles, because the drivers differ. If you are moving sector, prepare a clear account of what drives their revenue and cost base to close the gap.
How can I tell whether my modelling holds up under time pressure?
Run a timed, scored simulation before the real thing rather than another untimed practice build. Cohesyve lets you import the job you are applying for and run three free scored assessments, with a report showing where the clock, not your knowledge, was the constraint.
Practice
What a Financial Analyst assessment tests, and how it's scored →
Career guide
What a Financial Analyst actually does, and where the path leads →
Prepare for another role
Cohesyve · Practice for candidates
Don't walk in guessing. Find out where you stand first.
Three free scored practice runs, a full report on every one, and a learning pathway built from what you actually got wrong. No card required.
3
free assessments
$0
no card required
Full
scored report every run
Every
question type included
3 free assessments · No card required · Pro from $16/mo
For hiring teams
Hiring for a Financial Analyst role? See how your applicants perform before you spend interview time.